In communications, especially Long Term Evolution (LTE) cellular communications, when a communication device is transmitting in the uplink, the uplink signal often leaks into the downlink frequency band. While the power of the leaked signal is small as compared with the power of the transmit signal, the leaked signal power is still significant enough to prevent the base station from scheduling downlink transmissions on the full downlink frequency band. This is a waste of resources.
To alleviate this so-called self-interference, a cancellation algorithm may estimate self-interference from the transmitter at the receiver, and then cancel the estimated self-interference. This estimation, typically accomplished by adapting a filter, is challenging as the mapping from transmitted signal to interference is highly nonlinear and may have memory effect.